In the c-myc/TGF-a transgenic mouse model, hepatocellular adenoma (HCA) appear in the liver by 4 months of age and progress to hepatocellular carcinoma (HCC) by 8-10 months of age. Histologically, a small nest of HCC are frequently observed in HCA nodules, suggesting that HCA is a precancerous lesion in this model. In this study we have used cDNA microarray to analyze the difference in global gene expressions between HCA and HCC generated in these mice. One hundred nine genes out of 2070 genes (5.3%) were commonly up- or down-regulated more than 2-fold in at least 3 out of 10 comparisons, and no significant difference was shown in 1536 genes (74.2%) between the HCA and HCC samples. Expression of Igfbp1, Apoa4, Tfe2a, CD24a, Cyp2e1, Afp, Glns, Itga3, Cbr1, Pigf, and two EST were up-regulated, and expression of Car3, Cyp2j5, Ttf3, HMGCS2, Gnmt, Cyp2b9, Ptpn11, Lasp1, and two EST were down-regulated in at least 6 out of 10 comparisons. Hierarchical clustering analysis of the gene expression suggest that these 2 histologically diagnosed HCCs are close to HCAs in molecular level, and gene expression profile of a histologically diagnosed HCA is close to that of HCC. To validate our findings, five genes (Car3, TIEG, Afp, and Igfbp1) were selected for semi-quantitative RT-PCR analysis, and the RT-PCR results were consistent with the observation in cDNA microarray studies. One of the most frequently up-regulated genes in carcinoma was Angiogenin, and other genes such as VEGF, EGF, and FGF did not show a significant change. Furthermore, an intratumoral micro-vessel density by CD31 immunohistochemistry showed significantly increased angiogenesis in HCC as compared to the HCA (p<0.005), suggesting that Angiogenin has an important role in angiogenesis during hepatocellular carcinogenesis in this model. The data demonstrated that the application of unbiased analytical approaches to the gene expression profiles of mouse tumors is useful to distinguish HCC from HCA. In addition, a set of predictor genes for the classification of tumors during development of HCC in this model was identified based on gene expression profiles of HCA and HCC. We are also currently using the cDAN microarray technique to understand the molecular mechanisms of tumor metastasis and to identify molecular targets for therapy, it is essential to characterize genes that orchestrate the acquisition of metastatic conversion. Using cDNA microarrays containing 7K rat cDNAs, and 45K mouse cDNAs, we have conducted a systematic characterization of gene expression in v-raf or v-raf/v-myc transformed rat liver epithelial (RLE) cell lines that showed different degrees of metastasis. In the first array experiments, 7 transformed cell lines were compared with a parental RLE cell. The hierarchical clustering analysis of gene expression profiles revealed two groups characterized by their in vivo metastatic potential. Most strikingly, one non-metastatic cell line, RJ3611-T1lacZ (T1), was co-clustered with metastatic cell lines, suggesting that T1 underwent significant genetic changes, but still lacked the critical genetic elements for metastatic conversion. Cell lines in metastatic group including T1 expressed a low level of extracellular matrix proteins and cellular adhesion proteins, indicating that loss of extracellular matrix and cell-to-cell interaction appear to be first-rate limiting step for metastatic conversion. However, co-clustering of T1 with metastatic cell lines indicated that the loss of cell-to-cell interaction might not be sufficient for complete metastatic conversion. In the second array experiments, the non-metastatic T1 was compared against all metastatic cell lines in order to reveal the critical genes required for metastatic conversion but not expressed in the T1 cell line. Genes involved in immune cell homing such as osteopontin and galectin-3 were up-regulated in all metastatic cell lines, strongly suggesting that invasion through homing molecules is critical for metastasis. Most importantly, expression of anti-apoptotic genes and pro-apoptotic genes was up- and down-regulated respectively in all metastatic cell lines. Survival of intravasated cells in circulation system or extravasated cells in new environment might play a key role for metastasis. Our study provides a gene expression signature consistent with two critical events in the metastatic process, namely, the acquisition of homing capacity and increased survival potential of the metastatic cells.